Xfinity modem: XG2v2-P
After updating, I removed the Xfinity device and then added it back to make sure I have the latest codebase but I still have the same unreliability as before (buttons need to be pressed repeatedly before the STB responds).
I also have a second device configuration for the same Xfinity cable modem that uses learned codes from the original remote. This configuration generally works better than the Sofabaton device configuration, but isn't as reliable as the Logitech for example.
In an attempt to help try to get visibility into this issue, I recorded the IR signals as audio on my Mac using Sound Studio and a 192KHz sample rate. While not as good as an oscilloscope, it did show how the codes sendtby Sofabaton differed from the original in the first word. I have the audio recordings here:
Xfinity: https://sofabaton-ir-samples.s3.amazonaws.com/xfinity-power-single.m4a
Sofabaton: https://sofabaton-ir-samples.s3.amazonaws.com/sofabaton-power-single.m4a
I created images of these audio records by loading them as tracks in GarageBand and aligning them. For better resolution, I created an image for each of the two words (or what I call words):
First word is here: https://sofabaton-ir-samples.s3.amazonaws.com/xfinity-sofabaton-first-word.png
Second word is here: https://sofabaton-ir-samples.s3.amazonaws.com/xfinity-sofabaton-second-word.png
It's clear that the first word does not have the same pattern between the original remote and the Sofabaton remote. The second word has an identical pattern.
Looking more closely at the first part of the first word, it looks like it's missing a "spike" or "bit" at the beginning. In other words, it's appears to be short by 1 spike or bit.
Alas, I don't have a solution to offer. All I can hope for is that this post shines some (non-infrared) light on the issue so that it may get resolved sooner.
Marcel
